developing a seismic velocity model of the central valley ... · noise and group velocity data, and...
TRANSCRIPT
Award Number: G16AP00094
Developing a Seismic Velocity Model of the Central Valley,Northern California
Justin Lindeman, Donna Eberhart-Phillips, Louise H. Kellogg, and Lorraine J. Hwang
March 1, 2016 - February 28, 2017
University of California, Davis1 Shields Ave
Earth and Planetary SciencesDavis, CA 95616Fax: 530.752.0951
530.752.0350 [email protected] [email protected]
530.752.3690 [email protected] [email protected]
AbstractIn this report, we present five seismic velocity models for Northern California with em-
phasis on the Sacramento-San Joaquin Delta region (SSJD). A regional 3-D tomographicmodel (SSJD2016) of P-wave velocity (Vp) and Vp/Vs was calculated from a joint inversionof travel-time and group velocity observations using the Simul velocity inversion code. Cal-culated gravity measurements compared to gravity observations helped improve areas thatlacked seismic data. SSJD2016 extends from the Great Valley’s western margin to the Sierrawith a horizontal node spacing of 10 km and vertical node spacing of 3−8 km to 36 km depth.Compressional and shear-wave velocity versus depth relations were used to integrate Tertiarysubmarine canyons within the Cenozoic Great Valley sequence. The four shallow velocitymodels - SSJDOPHW95, SSJDGRANW95, SSJDFRANW95, and SSJDFRANG16, extendfrom the surface to 3.2 km. SSJDOPHW95, SSJDGRANW95, and SSJDFRANW95 utilizethe basement interpretations from Wentworth et al. [1995] and assume ophiolitic, granitic,and Franciscan rock, respectively, beneath the Great Valley sequence. SSJDFRANG16 ap-plies Graymer (written communication, 2016) and assumes only Franciscan rock to calculatevelocities beneath the Great Valley sequence.
1 Introduction
California’s San Andreas Fault System (SAFS) contains some of the most active faults inthe United States. Many of its segments which are prone to large and destructive earthquakestraverse heavily populated regions. The northern SAFS dissects the San Francisco Bay Area(SFBA) and extends ∼75 km eastward across the California Coast Ranges into the GreatValley and the Sacramento-San Joaquin Delta (SSJD). Recent activity (i.e. 2000 Mw 5.0Yountville and the 2014 Mw 6.0 Napa Valley earthquakes) on the West Napa fault raisesconcerns for a large (M>6) magnitude earthquake near and/or in the SSJD (Baltay and
Boatwright [2015]; Barnhart et al. [2015]; Boatwright et al. [2015]; Brocher et al. [2015];
1
Dreger et al. [2015]; Langenheim et al. [2006]; DRMS [2006]; Wesling and Hanson [2008]).Faults of the San Joaquin fault zone and the Sacramento Basin (e.g., Dunnigan, Midland,Sweitzer, Willow Hills) and the larger faults bordering the SSJD (e.g., Concord-Green Valley,Hayward-Rodgers Creek, Calaveras) (MacKevett [1992]; Pepper and Johnson [1992]; Wong
[1992]) are poorly studied and are capable of generating aMw 6.7 earthquake or greater (Fieldet al. [2014]). The 1892 ML 6.5 Vacaville-Winters earthquake occurred on an unidentifiedfault in the region (Bennett [1987]). The damage described is comparable to the 1983Coalinga earthquake of similar magnitude (Mw 6.2) that caused $31 million in damage inthat sparsely populated region (Bennett [1987]). Some have speculated that the Midlandfault is the causative fault, however, little conclusive information exists.
At the confluence of the Sacramento and San Joaquin Rivers, the construction of nearly1800 km of levees over the last century have transformed the SSJD from a tidal marshinto an agricultural-rich region and freshwater source for much of California. In doing so,human activity has added levee failure to the list of seismic risks in the region. Additionally,subsidence at a rate of 2−5 cm/yr (Rojstaczer et al. [1991]; Sharma et al. [2016]) has resultedin elevating the heights of levees, increasing stresses along the levee banks and hence, theirpotential for failure. An earthquake would subject the already fragile levees to dynamicloading possibly leading to failure and flooding (Torres et al. [2000]; DRMS [2006]).
The SSJD generates $35 billion annually (DRMS [2006]). The Delta Risk ManagementStrategy (DRMS [2006]) estimated that levee failure induced flooding could cost between$6.7 and $56.3 billion and damage could extend into the Sacramento and Stockton areas.A salt water incursion due to levee failure would have devastating consequences for theState’s freshwater supply impacting agriculture, the delta islands, sloughs, and man-madechannels. The SSJD is also home to major water infrastructure - the Delta-Mendota Canaland the California Aqueduct; transportation infrastructure - two interstates and three statehighways; and power infrastructure such as generation, storage, and electrical transmissionlines and facilities.
Strong ground motion in the SSJD has the potential of wide-spread damage to buildingsand to infrastructure. Both soil saturation and weak soils are recognize as important sitecharacteristics resulting in soil failure and damage to a built environment. Soft sedimentscompound this problem by amplifying seismic waves. The upper ∼100 m of the SSJD iscomposed of Holocene peats and unconsolidated sediments eroded from the Coast Ranges.The geometry of these deposits can further distort ground motions at soft rock sites andis important for wave-focusing effects. In addition, reverberation of seismic waves at basinedges can lead to interference patterns that could enhance the ground motion at longerperiods.
The Great Valley’s flat surface hides a complex subsurface that has been explored foroil, gas, and water. Its depth sharply increases westward under the SSJD extending possiblyto depths of 10 km or greater (Wentworth et al. [1995]; Moores et al. [2002]; Jachens et al.
[2006]; Thurber et al. [2007]). This deep structure impacts the lower frequency seismicwaves potentially causing more widespread effects. Steep edges of the basin could easilygenerate complex surface waves. Seismic waves traveling at high velocities through crystallinerock refract and slow dramatically when encountering the soft rocks of the basin. Thisincreases the amplitude of the earthquake waves and traps energy in the basin extending the
2
−124˚ −123˚ −122˚ −121˚ −120˚ −119˚
36˚
37˚
38˚
39˚
40˚
0 100
Kilometer
−124˚ −123˚ −122˚ −121˚ −120˚ −119˚
36˚
37˚
38˚
39˚
40˚
Sierra N
evada
Great V
alley
Coast R
anges
Sacramento
Napa
Stockton
Figure 1: Regional map of the study area in northern California showing active Quaternaryfaults (red), selected regions (blue), and selected cities (orange).
3
−124˚ −123˚ −122˚ −121˚ −120˚ −119˚
36˚
37˚
38˚
39˚
40˚
0 100
Kilometer
−124˚ −123˚ −122˚ −121˚ −120˚ −119˚
36˚
37˚
38˚
39˚
40˚
LegendUSGS Detailed Velocity ModelUSGS Regional Velocity ModelSeismic Velocity ModelShallow Velocity Model
Figure 2: Map showing the regions of the northern California velocity models in this study.Blue box shows the area of the USGS 08.03.1 detailed velocity model, dashed blue box theUSGS 08.03.1 regional velocity model, and yellow box the region of the shallow velocitymodel study. The red box shows the study area of SSJD2016. Velocities were held fixed inthe region west of the red dashed line.
4
duration of shaking. Hence, realistic models of the subsurface are needed for understandingthe variations in strong ground motion in order to estimate damage patterns from futureearthquakes and simulate phenomena affecting human safety.
This study improves the existing regional seismic velocity models by: (1) updating ex-isting seismic tomography models by first using recent earthquake data as well as ambientnoise and group velocity data, and then integrating land-based gravity measurements; and(2) developing shallow crustal velocity models by employing velocity-versus-depth relationswith two basement interpretations and introducing three major regional buried submarinecanyons. These models provide areas of higher resolution volume characteristics (e.g. seis-mic velocity and attenuation), geometries of structures, and boundary conditions useful forseismic-wave propagation modeling.
1.1 Previous Work
The most recent iterations of the USGS 3-D velocity models, version 08.3.0 (detailed andregional), developed for recent California ground motion studies (Aagaard et al. [2010a, b, 2013])were constructed using a broad range of geological and geophysical data including:
• geologic and geophysical mapping (Jachens et al. [2006]);
• lithology to: density, seismic compressional- and shear-wave speeds, and attenuationrelationships (Brocher [2005a, b, c, 2008]);
• waveform modeling (Rodgers et al. [2008]), and
• seismic tomography (Thurber et al. [2007]).
These studies are described below.
Jachens et al. [2006] produced a three-dimensional geologic map of northern Californiathat consists of specific geologic units separated by discrete boundaries. Their structuralmodel was derived from geologic mapping, gravity and magnetic surveying, double-differencerelocated seismicity, seismic soundings, seismic tomography, and well logs (Jachens et al.
[2006]). Depth to crystalline basement is taken from Wentworth et al. [1995] (Fig. 6) wherethe basement is defined in the eastern Great Valley by well logs (> 500), some of whichreached basement, and in the western valley, by a much smaller number of seismic refractionand reflection profiles. This 3-D block model provided structures where empirical relations(seismic-wave velocities, intrinsic attenuations, and densities) based on geology and depthcould be assigned.
Brocher [2005a, b, c] calculated velocity-depth relations for different lithologic units byfitting regression curves to available observations of borehole, seismic refraction and tomog-raphy, and density measurements for each main northern California rock type: Holocenesedimentary deposits; Plio-Quaternary sedimentary deposits; older Cenozoic sedimentaryrocks; Mesozoic Great Valley sequence sedimentary rocks; metagreywacke (Franciscan Com-plex); and granitic rocks (Brocher [2005a, b, c, 2008]). Quaternary-Tertiary regressions werederived from sonic logs for several basins, including the Great Valley and the western SSJDBrocher [2005c]. The Great Valley sequence (GRV) regressions are derived from a larger
5
sonic log collection and seismic refraction studies that suggest older GRV rocks (Mesozoicand Cenozoic) have slightly higher velocities than the Quaternary-Tertiary dataset. Later,Brocher [2008] calculated attenuation, Q, as a function of Vp and Vs. Qs and Qp were derivedfrom forward modeling of strong ground motions during the 1989 Loma Prieta and 1906 SanFrancisco earthquakes by Brocher [2008]. These relations were applied to the 3D-geologicmap of Jachens et al. [2006] to produce the USGS 3-D Seismic Velocity Model 05.1.0.
Rodgers et al. [2008] tested the validity of the USGS 05.1.0 model by comparing three-component broadband waveforms of moderate earthquakes (Mw 4 to 5) with synthetic seis-mograms calculated from USGS 05.1.0. Results showed predicted waves speeds were ∼5%faster than observed.
Coevally, Thurber et al. [2007] used data from the Northern California Earthquake DataCenter (NCEDC) that included phase arrival times from > 5500 chemical blasts and arrivaltimes from > 6000 earthquakes to develop a regional 3-D seismic velocity model for the SanFrancisco Bay Area (SFBA2007) using double-difference seismic tomography.
The USGS 3-D velocity model 08.3.0 combined the findings and models of Jachens et al.[2006], Brocher [2008], Rodgers et al. [2008], and Thurber et al. [2007] to ground-motionmodeling of earthquakes along the Hayward fault (Aagaard et al. [2010a, b]). Aagaard
et al. [2010a, b] improved the Brocher [2008] velocity-vs-depth relations to minimize theinconsistencies between the synthetic waveforms calculated with the 05.1.0 model and theobserved earthquake waveforms. In their study, calculated wave speeds were comparedfor each geologic unit in the 05.1.0 model with wave speeds in the SFBA2007 model. Therelations were improved to fit the gradients in the SFBA2007 model below several kilometersand the Brocher [2008] equations at depths closer to the surface. A detailed explanation canbe found in Aagaard et al. [2010b].
1.2 Data Used from Previous Studies
This study used three previously-collected datasets to solve for the velocity structure inthe SSJD. We integrated the Thurber et al. [2009] velocity model with the group velocitydataset from Fletcher et al. [2016] in a joint inversion. To construct a geologically-basedvelocity model, we used the Downey and Clinkenbeard [2010] structural dataset under theSSJD and applied velocity-vs-depth relations.
Thurber et al. [2009] expanded the SFBA2007 model creating a regional 3-D P-wavevelocity model of northern California (NC2009). Thurber et al. [2009] used over 5600 earth-quakes of magnitude 2.0 and above observed at 688 stations obtained from the NorthernCalifornia Earthquake Data Center (NCEDC). Their model has horizontal node-spacing of10 − 20 km and 3 − 8 km vertically and used finite-difference methods to compute traveltimes for over 5600 earthquakes and controlled-source experiments. The NC2009 velocitymodel revealed important regional features including a previously modeled high velocityophiolite body underlying the Great Valley (Godfrey et al. [1997]; Godfrey and Klemperer
[1998]; Moores et al. [2002]) and a never-before imaged deeply-penetrating fault beneaththe SSJD identified as the Pittsburg/Kirby Hills Fault. NC2009 was used as our startingvelocity model, however, NC2009 has poor resolution at shallow depths in the SSJD region.Lin et al. [2010] calculated a 3-D statewide velocity model that combined the NC2009 modelwith the Lin et al. [2007] southern California model. The Lin et al. [2010] model resulted in
6
a better velocity structure at the boundaries of these previous studies. Where NC2009 hadno representative velocities in the SSJD, Lin et al. [2010] was used as our initial model.
To this dataset we added the group velocities of Fletcher et al. [2016]. Fletcher et al. [2016]studied ambient noise from their temporary array in the SSJD (Fletcher and Boatwright
[2013]) to build a Rayleigh-wave group velocity model. In large parts of the SSJD, this datahas better resolution than the seismic tomography. This group velocity data was added tothe travel time data and jointly inverted as described in the next sections.
To refine the shallow velocity structure, we usedDowney and Clinkenbeard [2010]. Downeyand Clinkenbeard [2010] investigated the CO2 sequestration potential under the SSJD andidentified gross sandstone thickness to depths of ∼3 km. The depth-to-sandstone maps andgross sandstone isopachs were used to develop a model of the shallow structure by provid-ing structural constraints for three submarine canyons beneath the SSJD. These featuresare younger than the surrounding rocks and were assigned different Aagaard et al. [2010b]velocity-vs-depth relations then the surrounding Great Valley Sequence based on age andgeology.
2 Methods
2.1 Joint Inversion for Regional-Scale Tomography
2.1.1 Joint Inversion of Group Velocity and Earthquake Travel-time Data
The Simul velocity inversion codes use travel times of local earthquakes and controlledsources in a simultaneous inversion for hypocenters, 3-D P-velocity, Vp, and Vp/Vs (Thurberand Eberhart-Phillips [1999]; Eberhart-Phillips and Reyners [2012]), solving for velocity ona 3-D grid of nodes with velocity linearly interpolated between nodes, and allowing flexiblegridding through linking and fixing of nodes. The velocity is obtained by damped leastsquares, using LU decomposition to obtain a solution that has few artifacts and stays closeto the initial model where there is little resolution. Earthquake differential times and receiverdifferential times may be included to increase resolution near earthquake hypocenters andat shallow depths near stations. Each travel-time residual is related to perturbations to thevelocity at grid points along the raypath. In areas where there are many crossing raypathsthe velocity will be well resolved spatially; where there are few raypaths velocity featureswill be imaged broadly. In particular, there may be vertical smearing at shallow depthstowards the station such that shallow velocity features may be weakly imaged without sharpgradients unless there are numerous earthquakes and stations or active source constraints.
The Simul code has since been updated for joint inversion of travel-time and groupvelocity observations (Eberhart-Phillips and Fry [2016]). Group velocity (GV) observationsare included using each point from the GV 2-D models, for all periods. The GV 2-D modelsare a standard method of combining the uneven spatial sampling of interstation dispersioncurves, and are thus appropriate as input observations for 3-D velocity inversion. For eachpoint, an observation quality can be assigned based on the number of good dispersion pathsthat contribute to that volume. Surface waves sample a crustal volume along the surfacepath which may be 15 to 60 km wide, for periods 6 to 20 sec. Thus they have large horizontalaveraging compared to vertical averaging for travel-time. The GV is related to both Vp andVs from the surface throughout crust, with longer periods sampling deeper. The shallow crust
7
is sampled by all the periods and will have the most information. The complicated samplingdescribed by sensitivity kernels (partial derivatives) can be computed using Herrmann’scomputer programs in seismology (CPS; Herrmann [2013]), as shown for 10 sec period inFigure 3. The 10 sec GV is strongest from 5 to 10 km depth for Vs, and 0 to 5 km depth forVp. The width sampled is related to wavelength, so longer wavelengths sample larger widthas well as greater depth. We use CPS srfker96 routines for computing the model GV and thepartial derivatives that are used to relate the GV observations to the 3-D velocity model.
The GV residuals need to be related to the 3-D grid of Vp and Vp/Vs nodes, which arefiner spaced than the GV observation points, and the fine spacing needs to be retained for thedetailed travel-time velocity model. We take a box around each GV surface wave observationpoint. To compute the GV residual for point A, we take a 1-D average of the 3-D model inthe box and use that in srfker96. Then the 1-D kernels are related to all the 3-D velocitynodes within the box, in an analogous manner to the travel-time residual being related to 3-Dnodes along its raypath; with the Vs kernels related to partial derivatives of the Vp and Vp/Vs
model parameters. Thus, the GV residuals can be fit through perturbations to any pointsin the box volume while simultaneously retaining sharp features of the travel-time velocitymodel which typically may have 5 km grid spacing. Given the larger volume sampled by theGV, the 1-D velocity model specifically at point A need not fit the GV observation and the3-D model need not be as smooth as the GV 2-D observations. The box size is related to aninput parameter and the period of the GV observation. The box depth for a given periodGV observation extends to the depth where the Vs kernel is less than 4% of the maximumVs kernel. Then the 3-D velocity model can be updated by simultaneous inversion of theGV observations and travel-time observations using iterative damped least squares. At eachiteration, the residuals and partial derivatives are recalculated.
2.1.2 Gravity Inversion for 3D Seismic Inversion
Gravity observations also sample the 3-D volume and can aid in areas with little seismicdata. The calculated gravity from the 3-D model is obtained using the density-velocity rela-tionship of Brocher [2005b]. Both the observed and calculated gravity are upward continued2 km to de-emphasize near-surface density variations that have finer scale than the velocitygrid. The isostatic residual gravity is used because it is related to crustal heterogeneity.Each gravity observation is related to a large volume with the width of influence increasingwith depth. The gravity residuals are used in an inversion for 3-D velocity perturbations(Eberhart-Phillips and Michael [1993]; Eberhart-Phillips and Bannister [2002]). This is nota joint inversion method and the seismic velocity can be influenced by many factors otherthan density, such as fluid pressure and crack density.
2.1.3 Data
For this study, we used travel-time data from 281 earthquakes with magnitudes between2.0 − 5.2 during 2005-2016 recorded by the NCSN (NCDEC [2014]) which provided 49,215observations. Additionally, we incorporated travel-time data from refraction profiles thatwere used by Thurber et al. [2009] which provided an additional 19,189 observations. Surfacewave GV observations were obtained from Fletcher et al. [2016]. They used ambient noiseinterstation dispersion curves of GV to obtain 2-D models of GV at periods of 4.5 to 10.5s. These will be best constrained where the location is sampled by multiple interstationpaths, and hence, we assign observation quality based on the number of paths near each GV
8
point for 3273 GV observations. Gravity data was provided by Victoria Langenheim, USGS(written communication, 2016), and consisted of gridded observations of isostatic residualswith 2 km spacing and upward continued 2 km.
2.1.4 Inversion Procedure
We carried out a series of three velocity inversions, as described above, using a modifiedNC2009 as our starting model. NC2009 modeled the regional structure along a 10x20 kmgrid. Along paths where there was no data, the model could not be updated and stayedas the initial northern California 1-D model. Thus, many shallow parts of the SSJD lackrepresentative velocities. Hence, for our initial model we modified NC2009 using Lin et al.
[2010] model for depths -1 to 4 km in the SSJD.
While the group velocity data samples the whole crust, it provides the most improvementto the NC2009 model in the shallow crust. In the first inversion, the GV and travel-timeseismic data were used to update the 3-D velocity model at depths of −1 to 4 km, withthe deeper model fixed. In a second seismic inversion, the solution included depths of −1to 36 km, obtaining a seismic model that fits both seismic data sets. This decreased thedata variance of the P travel-time by 24.1%, the S-P travel-time by 14.2%, and the GVby 69.5% from the initial model. Finally, the seismic model was used as the initial modelfor an inversion of the isostatic residual gravity to obtain a seismic velocity model that isalso consistent with gravity. The final model, SSJD2016, provided a 98.8% decrease in datavariance of the gravity data while somewhat weakening the fit to seismic data. Overall theseismic-gravity 3-D inversions decreased the data variance of the P travel-time by 16.6%, theS-P travel-time by 13.1%, and the GV by 64.0%.
The results for depths of 1 km and 4 km are shown in Figure 4. The area outside theSSJD is included but was fixed to NC2009. An example of the GV at the 8.5 s periodobservation points calculated from the final 3-D velocity model is shown in Figure 5. Thelargest GV residuals tend to be in areas where there is competition with refraction data, forexample, in the southwest.
2.2 Shallow Velocity Models
2.2.1 Data
The area of the shallow velocity model is defined by the study area of Downey and
Clinkenbeard [2010]. Their study identified thick sedimentary sections suitable for carbonsequestration in California. Amongst these, basins in the Sacramento and San JoaquinValleys were found to have some of the thickness formation sequences mainly composed ofthe Cretaceous deposits of the Mokelumne River (MRF), Starkey (SF), and Winters (WF)formations. Regionally, these formations are incised and filled with the Tertiary sedimentsof the Markley, Martinez, and Meganos submarine canyons. Using information from over6200 gas well logs, Downey and Clinkenbeard [2010] identified thick sandstone sequences ofup to 460 m within the MRF, SF, and WF and developed isopachs of gross sandstone andsealing formation thickness along with their depths. Also identified were regions in whichsubmarine canyons incised each formation. The horizontal boundaries of these submarinecanyons along with the gross sandstone isopachs are used to estimate the thickness of thecanyons through each formation. Since all canyons cut the MRF, the top of the canyons aretaken to be the depth to the top of the MRF sandstones. The thickness of each canyon are
9
Figure_EP_1
Figure 3: Sensitivity kernels at T=10 sec using Herrmann [2013] srfker96, Vp (red), Vs (blue),and Vp/Vs (green).
10
2.53.03.54.04.55.05.56.06.57.07.58.08.59.09.5
Vp (km/s)
X-Distance (km X-Distance (km
-300
-275
-250
-225
-200
-175
-150
-125
-100
-75
-50
-25
0
25
Y-D
ista
nce
(km
)
-100-50050100
Z= 1 vpgi2sdv34
-122
-121
-120
39
38
37
3.5
3.5
4
4
4.5
4.5
555
5
5.5
5.5
6
6
6
6
6.5
-300
-275
-250
-225
-200
-175
-150
-125
-100
-75
-50
-25
0
25
-100-50050100
Z= 1 vpvssdv34
-122
-121
-120
39
38
37
1.75
1.7
5
1.75
1.75
1.775
1.7
75
1.7
75
1.775
1.8
1.8
1.8
1.60 1.65 1.70 1.75 1.80 1.85 1.90 1.95
Vp/Vs
-300
-275
-250
-225
-200
-175
-150
-125
-100
-75
-50
-25
0
25
Y-D
ista
nce
(km
)
-100-50050100
Z= 4 vpgi2sdv34
-122
-121
-120
39
38
37
4
4.5
4.5
5
5
5
5.5
5.5
5.5
6
6
6.5
6.5
-300
-275
-250
-225
-200
-175
-150
-125
-100
-75
-50
-25
0
25
-100-50050100
Z= 4 vpvssdv34
-122
-121
-120
39
38
37
1.7
25
1.75
1.75
1.7
5
1.7
75
1.7
75
1.7
75
1.77
5
1.775
1.8
1.8
1.8
1.8
Figure_EP_2
Figure 4: Depth slices of SSJD2016 showing Vp (left) and Vp/Vs (right) at depths = 1 and 4km, with sources (earthquakes and refraction shots) used in the inversion shown by x’s. Thearea outside the Delta is included but was fixed to NC2009; the inversion area is boundedby x= 80 to -120, y= -300 to 40, and z= -1 to 36. Coastline, active faults, and Sacramentoand San Joaquin Rivers are shown. Stars are Sacramento and Stockton, solid diamonds areMt. Diablo, Mt. Hamilton, and Mt. Whitney.
11
a) calculated group velocity238 239
122 121 120
240
37 37
38 38
39 39
1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8
GV km/s
b) residual group velocity238 239 240
37 37
38 38
39 39
-0.4 -0.2 0.0 0.2 0.4
GVresid km/s
122 121 120
Figure 5: Plot of the a) calculated group velocity and b) the residual group velocity atthe group velocity observation points which are the locations where the 2-D group velocity(Fletcher et al. [2016]) included multiple dispersion measurements.
assumed to extend to the bottom of the last formation in which that canyon appears. TheMarkley submarine canyon is present from the top of the MRF to the bottom of the SF. TheMartinez and Meganos submarine canyons are only present in the MRF, but appear lowerthan the Markley in the west because of the Midland fault offset.
2.2.2 Calculating Velocity with Geology
The shallow velocity models presented in this study (SSJDOPHW95, SSJDGRANW95,SSJDFRANW95, and SSJDFRANG16) apply the compressional (Vp) and shear-wave (Vs)seismic velocity versus depth relations and intrinsic attenuation for common rock types innorthern California originally presented by Brocher [2008] and later modified by Aagaard
et al. [2010a, b]. The equations in Table 1 are a subset of the Aagaard et al. [2010b] equationscurrently used in the USGS 08.3.0 Seismic Velocity Model and are used to calculate Vp, Vs,Qs, and Qp in the shallow velocity models presented here. For equations that are at thesurface, the velocity is assumed. For example, we assume all rocks outside the submarinecanyons are Great Valley sequence (GRV). Equation 7 solves for GRV Vp in the range of 2.5km/s (at surface; z = 0) to 4.36 km/s (at 3.2 km; z = 3.2). Equations 8 and 9 solve forGRV Vs in the range of 0.6 km/s (at surface; z = 0) to 2.41 km/s (at 3.2 km; z = 3.2).Equation 21 solves for GRV Qs in the range of 39.17 (at surface; z = 0) to 203.17 (at 3.2km; z = 3.2). Equation 22 solves for GRV Qp in the range of 78.34 (at surface; z = 0) to406.34 (at 3.2 km; z = 3.2). Inside the submarine canyons, equations 1− 6 calculate Vp andVs with a range of 0.7− 4.1 and 0.08− 0.68 km/s, respectively.
It is unclear at what depth the transition from GRV to crystalline rock occurs or even
12
−125˚ −124˚ −123˚ −122˚ −121˚ −120˚
35˚
36˚
37˚
38˚
39˚
−125˚ −124˚ −123˚ −122˚ −121˚ −120˚
35˚
36˚
37˚
38˚
39˚
14
10
6
2LegendW95 Basement ContourG16 Basement ContourSeismic LinesWell Locations
3
45
5
3
5
8
2
9
8
6
8
1
Figure 6: Basement surface contours, well locations, and seismic lines from Wentworth et al.
[1995]. Basement is better constrained to the east where most of the wells are located. Insetadds the basement surface contours from Graymer (written communication, 2016).
13
what rock type underlies the GRV. To address this ambiguity, two different datasets areused to define basement under the SSJD. The first dataset from Wentworth et al. [1995]maps the crystalline basement from outcrops to the east in the Sierra Nevadan foothillsand then proceeding west along a steeply dipping interface. Their model does not reachto the western region of our shallow velocity models; the lower boundary of our shallowvelocity models are well above the interpreted basement surface in the west. The seconddataset from Graymer (written communication, 2016) used well logs, seismic refraction lines,and downward projection of surficial gravity measurements to interpret basement depthbeneath the SSJD to provide a detailed basement map in the western SSJD. In contrastto the Wentworth dataset, Graymer maps western basement shallower then 3.2 km underthe western SSJD. There are subtle disagreements in the two basement maps (Fig. 6) andoverlap between the two basements maps is minimal. In addition, the Graymer datasetconflicts with the carbon sequestration isopachs (see Discussion). For this last reason, wedid not attempt to combine the two basement maps.
As a result of the differences in the basement interpretations, four models have been pro-
−122˚ −121˚
38˚
39˚
−122˚ −121˚
38˚
39˚
3.2
3.6
A
A
A’
B
B’C
C’
D
D’
Vp, depth = 1.0 km
−122˚ −121˚
38˚
39˚
−122˚ −121˚
38˚
39˚
3.6
3.6
4
B
Vp, depth = 1.6 km
2.5 3.0 3.5 4.0 4.5 5.0 5.5
km/s
SSJD2016
Figure 7: Map-view slices of SSJD2016 at 1.0 and 1.6 km depths for the region of the shallowvelocity study. Contour interval is 0.2 km/s. White lines bound the three submarine canyonsas projected in the MRF depth-to-sandstone map (Downey and Clinkenbeard [2010]). Thelocation of cross-section slices in Fig. 8 are shown. The SSJD2016 locations correspond to:(1) A-A’, x = 0 km; (2) B-B’, y = -68 km; (3) C-C, x = -22 km; and (4) D-D’, y = -122 km.
14
duced: three models use theWentworth et al. [1995] basement interpretations calculated withthree different basement geologies ophiolitic (SSJDOPHW95), granitic (SSJDGRANW95),and Franciscan basement (SSJDFRANW95); and one model with the Graymer basementinterpretations calculated assuming Franciscan basement (SSJDFRANG16). The basementvelocities are calculated in one of three ways. First, we assume that ophiolitic material (asdefined by Aagaard et al. [2010b]) underlies the GRV and utilize the Great Valley and maficregressions (Table 1, eq. 19 and 20). Second, we assume granitic rock underlies the GRV andapply the granitic regressions (Table 1, eq. 14-18). Third, we assume Franciscan materialbeneath the GRV (Table 1, eq. 10-13).
Above the basement interface, the addition of submarine canyons creates slow anomalieswhich are 0.2 to 0.6 km/s slower than the surrounding GRV. Beneath the basement inter-face the Vp structure changes. SSJDOPHW95 is consistently 5.9 km/s from the basementinterface to the lower bounds of the model. SSJDGRANW95 ranges from 4.65 to 5.66 km/sbeneath the basement. SSJDFRANW95 basement velocities range from 4.47 to 5.37 km/s.
0
1
2
3Depth
(km
)
−40 −30 −20 −10
Distance (km)
3.23.2
3.63.6
D D’
2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
km/s
0
1
2
3Depth
(km
)
−140 −130 −120 −110 −100 −90 −80 −70 −60
2.8
3.2
3.2
3.6
C’ C
0
1
2
3Depth
(km
)
−40 −30 −20 −10 0 10
3.2
3.6
4 4.4
B B’
0
1
2
3Depth
(km
)
−100 −90 −80 −70 −60 −50 −40
3.2
3.6
4
A’ A
Figure 8: Cross-section slices through SSJD2016 showing the shallow velocity structure. See(Fig. 7) for location of cross sections.
15
Table 1: Summary of Vp and Vs vs Depth Relations
RelationDepth (z)
Range*Rock Type
1. Vp = 0.7 + 42.968z − 575.8z2 + 2931.6z3 − 3977.6z4 0− 0.04 Quat.-Tertiary sed.
2. Vp = 1.5 + 3.735z − 3.543z2 0.04− 0.5 Quat.-Tertiary sed.
3. Vp = 2.24 + 0.6z 0.5− 10 Quat.-Tertiary sed.
4. Vs = 0.08 + 2.5z 0− 0.025 Quat.-Tertiary sed.
5. Vs = (Vp − 1.36)/1.16 0.025− 0.05 Quat.-Tertiary sed.
6. Vs = 0.7858− 1.2344Vp + 0.7949V 2
p− 0.1238V 3
p+ 0.00064V 4
p≥ 0.05 Quat.-Tertiary sed.
7. Vp = 2.5 + 0.5833z 0− 3.6 Great Valley sequence
8. Vs = 0.6 + 1.183z 0− 1 Great Valley sequence
9. Vs = 1.5 + 0.2836z 1− 5 Great Valley sequence
10. Vp = −0.03 + 2.5 + 2z 0− 1 Franciscan
11. Vp = −0.03 + 4.5 + 0.45(z − 1) 1− 3 Franciscan
12. Vp = −0.03 + 5.4 + 0.00588(z − 3) ≥ 3 Franciscan
13. Vs = 0.7858− 1.2344Vp +0.7949V 2
p− 0.1238V 3
p+0.00064V 4
p≥ 0 Franciscan
14. Vp = 1.5 + 5z 0− 0.5 Granitic rocks
15. Vp = 4 + 1.3(d− 0.5) 0.5− 1.5 Granitic rocks
16. Vp = 5.3 + 0.3(d− 1.5) 1.5− 2.5 Granitic rocks
17. Vp = 5.6 + 0.08(z − 2.5) 2.5− 5 Granitic rocks
18. Vs = 0.7858− 1.2344Vp +0.7949V 2
p− 0.1238V 3
p+0.00064V 4
p≥ 0 Granitic rocks
19. Vp = 5.9 0− 10 Mafic & GV ophiolite
20. Vs = 0.7858− 1.2344Vp +0.7949V 2
p− 0.1238V 3
p+0.00064V 4
p≥ 0 Mafic & GV ophiolite
21. Qs = −16 + 104.13Vs − 25.225V 2
s+ 8.2184V 3
s≥ 0 All rocks
22. Qp = 2Qs ≥ 0 All rocks
*Vp in km/s; Depth in km.From Aagaard et al. [2010b]
16
−122˚ −121˚
38˚
39˚
−122˚ −121˚
38˚
39˚
E
A
A’
B
B’C
C’
D
D’
Vp, depth = 1.0 km
−122˚ −121˚
38˚
39˚
−122˚ −121˚
38˚
39˚
F
Vp, depth = 1.6 km
2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
km/s
SSJDFRANW95
38˚
39˚
38˚
39˚
C
Vp, depth = 1.0 km
38˚
39˚
38˚
39˚
D
Vp, depth = 1.6 kmSSJDGRANW95
−122˚ −121˚
38˚
39˚
−122˚ −121˚
38˚
39˚
A
Vp, depth = 1.0 km −122˚ −121˚
38˚
39˚
−122˚ −121˚
38˚
39˚
B
Vp, depth = 1.6 kmSSJDPHW95
Figure 9: Map-view slices of A and B: SSJDOPHW95 (ophiolite), C and D: SSJDGRANW95 (granite), and E and F: SSJDFRANW95 (Franciscan) at depths of left: 1.0 km andright: 1.6 km depths. White lines bound the three submarine canyons as projected in theMRF depth-to-sandstone map (Downey and Clinkenbeard [2010]). See (Fig. 7) for locationof cross sections.
17
0
1
2
3Depth
(km
)
−40 −30 −20 −10
Distance (km)
2.8
3.23.2
3.6
4 5.2
D D’
2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
km/s
0
1
2
3Depth
(km
)
−140 −130 −120 −110 −100 −90 −80 −70 −60
2.82.8
3.23.2
3.23.2
3.6
44
C’ C
0
1
2
3Depth
(km
)
−40 −30 −20 −10 0 10
2.82.8
3.23.2
3.6
4
B B’
0
1
2
3Depth
(km
)
−100 −90 −80 −70 −60 −50 −40
2.82.8
3.23.2
3.63.6
44
A’ A
Figure 10: Cross-section slices through SSJDFRANW95 along: (1) A-A’, (2) B-B’, (3) C-C’,and (4) D-D’. Contour interval is 0.2 km/s.
18
3 ResultsThe final map-view slices and cross sections for seismic model SSJD2016 are shown
in Figure 4. These are cropped to the shallow velocity model area in Figures 7 and 8.Modifications to the shallow subsurface velocities using the Wentworth-basement are shownin Figure 9. Cross sections for SSJDFRANW95 are shown in Figure 10 and are representativeof this group of models.
Model data has been archived and is briefly described below. Please see the archivedpackages for complete data descriptions.
3.1 SSJD2016 Velocity Model Data
filename: vlSSJD2016 2km xyzltln.txt
format: ascii
parameters: latitude, longitude, x, y, depth, Vp, Vp/Vs, Vs, DWS−Vp, and DRE−Vp
Comments: Figures of additional depth slices and cross sections are also part of thisdata package.
doi: 10.5281/zenodo.556605
3.2 Shallow Velocity Model
The shallow velocity model data files contains both the velocity model data and thescripts.
doi: 10.5281/zenodo.569057
format: ascii
Comments: Figures of additional depth slices and cross sections are part of this datapackage in addition to all report figures.
Script Files
filename: shallownmodelOPH.m
filename: shallownmodelGRAN.m
filename: shallownmodelFRAN.m
Data Files
filename: SSJDshallowmodelOPHW95.out
filename: SSJDshallowmodelGRANW95.out
filename: SSJDshallowmodelFRANW95.out
filename: SJDshallowmodelFRANG16.out
parameters: longitude, latitude, x, y, depth, Vp, Vs, Qs, and Qp.
19
−123˚ −122˚ −121˚ −120˚ −119˚36˚
37˚
38˚
39˚
40˚
−123˚ −122˚ −121˚ −120˚ −119˚36˚
37˚
38˚
39˚
40˚
−0.5
−0.5
0
0
0
0
0
0.5
0.5
E
Vp, depth = 20 km
−123˚ −122˚ −121˚ −120˚ −119˚36˚
37˚
38˚
39˚
40˚
−123˚ −122˚ −121˚ −120˚ −119˚36˚
37˚
38˚
39˚
40˚
0
0
0
0
0
0
0
0
0
F
Vp, depth = 28 km
−2.5 −2.0 −1.5 −1.0 −0.5 0.0 0.5 1.0 1.5 2.0 2.5
km/s
36˚
37˚
38˚
39˚
40˚
36˚
37˚
38˚
39˚
40˚
−0.5
−0.5
0
0
0
0
0
0.5
0.5
1
C
Vp, depth = 8 km
36˚
37˚
38˚
39˚
40˚
36˚
37˚
38˚
39˚
40˚
−0.5
−0.5
−0.5
−0.5
0
0
0
0
0
0
0.5
0.5
0.5
D
Vp, depth = 14 km
−123˚ −122˚ −121˚ −120˚ −119˚
36˚
37˚
38˚
39˚
40˚−123˚ −122˚ −121˚ −120˚ −119˚
36˚
37˚
38˚
39˚
40˚
−2
−2
−1.5
−1.5
−1
−1
−1
−1−0.5
−0.5
0
0
0
0
0
0.5
0.5
0.5
0.5
0.5
0.5
1
A
Vp, depth = 1 km−123˚ −122˚ −121˚ −120˚ −119˚
36˚
37˚
38˚
39˚
40˚−123˚ −122˚ −121˚ −120˚ −119˚
36˚
37˚
38˚
39˚
40˚
−1
−1
−0.5
−0.5
0
0
0
0
0.5
0.50.5
0.5
1
B
Vp, depth = 4 km
Figure 11: Map-view slices of residuals from SSJD2016 compared to NC2009 at A. 1.0 km,B. 4 km, C. 8 km, D. 14 km, E. 20 km, and F, 28 km depths. Positive residuals representslower SSJD2016 velocities relative to NC2009 velocities. Negative residuals represent fasterSSJD2016 velocities relative to NC2009. Contour interval is 0.25 km/s.
20
−122˚ −121˚
38˚
39˚
−122˚ −121˚
38˚
39˚
−0.8
−0.6
−0.4−0.2
0
0
0
0.2
0.4
0.4
0.6
E
Vp, depth = 1.0 km
−122˚ −121˚
38˚
39˚
−122˚ −121˚
38˚
39˚
−1
−0.8
−0.6
−0.4
0
0
0
0.2
0.20.2
0.4
0.8
F
Vp, depth = 1.6 km
−2.5 −2.0 −1.5 −1.0 −0.5 0.0 0.5 1.0 1.5 2.0 2.5
km/s
SSJDFRANW95
38˚
39˚
38˚
39˚
−0.8
−0.6
−0.4
0
0
0
0.2
0.4
0.4
0.6
C
Vp, depth = 1.0 km
38˚
39˚
38˚
39˚
−1.6
−1.4
−1.2
−1
0
0
0
0.2
0.20.2
0.4
0.8
D
Vp, depth = 1.6 kmSSJDGRANW95
−122˚ −121˚
38˚
39˚
−122˚ −121˚
38˚
39˚
−2.2
−2
−1.8
0
0
0
0.2
0.4
0.4
0.6
A
Vp, depth = 1.0 km −122˚ −121˚
38˚
39˚
−122˚ −121˚
38˚
39˚
−2.2
−2 −1.8
−1.6
0
0
0
0.2
0.20.2
0.4
0.8
B
Vp, depth = 1.6 kmSSJDPHW95
Figure 12: Map-view slices of residuals from SSJDOPHW95 (A and B), SSJDGRANW95(C and D), and SSJDFRANW95 (E and F) compared to SSJD2016 at 1.0 km (left) and 1.6km (right) depths. Contour interval is 0.1 km/s. Yellow lines bound the three submarinecanyons as projected in the MRF depth-to-sandstone map. Within these models, only theWentworth et al. [1995] basement interpretations are used.
21
4 Discussion
We improved upon NC2009 in the SSJD by incorporating new travel time data and GV,ambient noise, and gravity measurements. This improved upon the resolution of previousstudies especially in regions in which traditional travel time data did not sample in theshallow crust. The differences between NC2009 and SSJD2016 models are predominantlynear sea level (depth = 0 km). As depth increases, the differences in the model velocitiesbecome smaller (Fig. 11). At depth = 1 km, SSJD2016 is generally slower than NC2009west of the Sierra Nevada (SN) and faster (∼1-2.5 km/s) under the SN (Fig. 11A). In theSSJD, SSJD2016 is generally ∼1 km/s slower than NC2009. The most noticeable differenceexists near Stockton where the SSJD2016 is ∼2.6 km/s slower than NC2009. At depth = 4km, SSJD2016 is slower west of the SN and faster under the SN, but the discrepancies aresmaller (Fig. 11B). Under the SSJD, the Great Valley south of Stockton, the Coast Rangeswest of the SSJD, and just north of Sacramento near the Sutter Buttes, velocities are slowerby 0.5 − 1.3 km/s. Under the SN velocities in SSJD2016 can differ as much as 1.8 km/s.Below 4 km, the velocity differences between the two models decrease to 0.4 km/s (Fig. 11C-F).
It is well known that an ophiolite sequence underlies the Great Valley group in the CoastRanges and a magnetic, gravity, and seismic anomaly, possibly another ophiolitic body, existsunderneath the Great Valley (Bailey et al. [1970]; Cady [1975]; Dickinson et al. [1996]). TheGreat Valley Ophiolite (GVO) anomaly extends nearly the entire length of the Sacramentoand San Joaquin Valleys (Cady [1975]; Godfrey et al. [1997]; Godfrey and Klemperer [1998];Godfrey and Dilek [2000]; Moores et al. [2002]; Thurber et al. [2009]; Lin et al. [2010]), yetis never exposed in outcrop. Previous models (Godfrey et al. [1997]; Godfrey and Klem-
perer [1998]; Godfrey and Dilek [2000]) led to interpretations that in northern California theGVO is an unserpentinized mantle section and if present in the southern valley, has beenserpentinized to a degree that it cannot be distinguished from Sierran basement. Moores
et al. [2002] hypothesized that the GVO contains a crustal-mantle sequence that overliesmetamorphic belts and a second crustal-mantle sequence. Due to the uncertainty about thecomposition of the material beneath the Great Valley, we produced the four shallow velocitymodels each with a different basement rock type and/or basement interpretation. Compar-ison of the basement map of Graymer (personal communication, 2016) and the sandstoneisopachs of Downey and Clinkenbeard [2010] reveals discrepancies in the western region of theSSJD. In Graymer’s map, the basement is only present in the western region where the CoastRange ophiolite is known to be present. SSJDFRANG16 assumes Franciscan basement. TheGraymer basement is much shallower than the sandstone depths from Downey and Clinken-
beard [2010]. Graymer’s basement extends into the sandstone layers defined by Downey and
Clinkenbeard [2010]. We compared the four velocity models to the new tomography model(SSJD2016). The residuals between SSJD2016 and SSJDOPHW95, SSJDGRANW95, andSSJDFRANW95 are shown in Fig. 12. We favor SSJFRANW95 because it compares well tothe tomography (Fig. 12 E and F).
22
5 Acknowledgements
Waveform data, metadata, and data products for this study were accessed through theNorthern California Earthquake Data Center (NCEDC), doi:10.7932/NCEDC. We are grate-ful to Victoria Langenheim and Russ Graymer for providing gravity data and basementinterpretations in the Delta.
6 References
Aagaard, B. T., R. W. Graves, D. P. Schwartz, D. A. Ponce, and R. W. Graymer, Ground-Motion Modeling of Hayward Fault Scenario Earthquakes, Part I: Construction of the Suiteof Scenarios, Bulletin of the Seismological Society of America, 100 , 2927–2944, 2010a.
Aagaard, B. T., M. G. Knepley, and C. A. Williams, A Domain Decomposition Approach toImplementing Fault Slip in Finite-Element Models of Quasi-static and Dynamic CrustalDeformation, Journal of Geophysical Research: Solid Earth, 118 , 3059–3079, 2013.
Aagaard, B. T., R. W. Graves, A. Rodgers, T. M. Brocher, R. W. Simpson, D. Dreger, N. A.Petersson, S. C. Larsen, S. Ma, and R. C. Jachens, Ground-Motion Modeling of HaywardFault Scenario Earthquakes, Part II: Simulation of Long-Period and Broadband GroundMotions, Bulletin of the Seismological Society of America, 100 , 2945–2977, 2010b.
Bailey, E. H., M. Blake Jr, and D. L. Jones, On-land Mesozoic Oceanic Crust in CaliforniaCoast Ranges, Geological Survey Professional Paper , 770-C , C70–C81, 1970.
Baltay, A. S., and J. Boatwright, Ground-Motion Observations of the 2014 South NapaEarthquake, Seismological Research Letters , 86 , 355–360, 2015.
Barnhart, W. D., J. R. Murray, S. Yun, J. L. Svarc, S. V. Samsonov, E. J. Fielding, B. A.Brooks, and P. Milillo, Geodetic Constraints on the 2014 M 6.0 South Napa Earthquake,Seismological Research Letters , 86 , 335–343, 2015.
Bennett, J., Vacaville-Winters Earthquakes, 1892, California Geology , 40 , 75–83, 1987.
Boatwright, J., J. L. Blair, B. T. Aagaard, and K. Wallis, The Distribution of Red andYellow Tags in the City of Napa, Seismological Research Letters , 86 , 361–368, 2015.
Brocher, T. M., A Regional View of Urban Sedimentary Basins in Northern California Basedon Oil Industry Compressional-Wave Velocity and Density Logs, Bulletin of the Seismo-
logical Society of America, 95 , 2093–2114, 2005a.
Brocher, T. M., Empirical Relations Between Elastic Wavespeeds and Density in the Earth’sCrust, Bulletin of the Seismological Society of America, 95 , 2081–2092, 2005b.
Brocher, T. M., Compressional and Shear Wave Velocity Versus Depth in the San FranciscoBay Area, California: Rules for USGS Bay Area Velocity Model 05.0.0, Tech. rep., U.S.Geological Survey Open-File Report 05-1317, 2005c.
23
Brocher, T. M., Compressional and Shear-Wave Velocity versus Depth Relations for CommonRock Types in Northern California, Bulletin of the Seismological Society of America, 98 ,950–968, 2008.
Brocher, T. M., et al., The Mw 6.0 24 August 2014 South Napa Earthquake, Seismological
Research Letters , 86 , 309–326, 2015.
Cady, J. W., Magnetic and Gravity Anomalies in the Great Valley and Western SierraNevada Metamorphic Belt, California, Geological Society of America Special Papers , 168 ,1–56, 1975.
Dickinson, W. R., et al., Alternate Origins of the Coast Range Ophiolite (California): Intro-duction and Implications, GSA Today , 6 , 1–10, 1996.
Downey, C., and J. Clinkenbeard, Preliminary Geologic Assessment of the Carbon 800 Se-questration Potential of the Upper Cretaceous Mokelumne River, Starkey, and WintersFormations - Southern Sacramento Basin, PIER Collaborative Report , California Energy
Commission, Sacramento, CA, 2010.
Dreger, D. S., M. Huang, A. Rodgers, T. Taira, and K. Wooddell, Kinematic Finite-SourceModel for the 24 August 2014 South Napa, California, Earthquake from Joint Inversionof Seismic, GPS, and InSAR Data, Seismological Research Letters , 86 , 327–334, 2015.
DRMS, Delta Risk Management Strategy: Probabilistic Seismic Hazard Analysis for GroundShaking and Estimation of Earthquake Scenario Probabilitiees, Tech. rep., URS Corpora-tion/Jack R. Benjamin & Associates, Inc., 2006.
Eberhart-Phillips, D., and S. Bannister, Three-dimensional Crustal Structure in the SouthernAlps Region of New Zealand from Inversion of Local Earthquake and Active Source Data,Journal of Geophysical Research: Solid Earth, 107 , ESE 15–1–ESE 15–20, 2002.
Eberhart-Phillips, D., and B. Fry, Joint Surface Wave and Earthquake Travel-Time Inver-sion for Western North Island, New Zealand, 3-D Velocity Model, Seismological Research
Letters , 87 , 546, 2016.
Eberhart-Phillips, D., and A. J. Michael, Three-dimensional Velocity Structure, Seismicity,and Fault Structure in the Parkfield Region, Central California, Journal of Geophysical
Research: Solid Earth, 98 , 15,737–15,758, 1993.
Eberhart-Phillips, D., and M. Reyners, Imaging the Hikurangi Plate Interface Region, withImproved Local-earthquake Tomography, Geophysical Journal International , 190 , 1221–1242, 2012.
Field, E. H., T. E. Dawson, K. R. Felzer, A. D. Frankel, V. Gupta, T. H. Jordan, T. Parsons,M. D. Petersen, R. S. Stein, R. J. Weldon, and others, Uniform California EarthquakeRupture Forecast, Version 3 (UCERF3)- The Time-Independent Model, Bulletin of the
Seismological Society of America, 104 , 1122–1180, 2014.
24
Fletcher, J. B., and J. Boatwright, Site Response and Basin Waves in the Sacramento-SanJoaquin Delta, California, Bulletin of the Seismological Society of America, 103 , 196–210,2013.
Fletcher, J. B., J. Erdem, K. Seats, and J. Lawrence, Tomographic Rayleigh-wave GroupVelocities in the Central Valley, California Centered on the Sacramento-San Joaquin Delta,Journal of Geophysical Research: Solid Earth, 103 , 2429–2446, 2016, 2015JB012376.
Godfrey, N. J., and Y. Dilek, Mesozoic Assimilation of Oceanic Crust and Island Arc into theNorth American Continental Margin in California and Nevada: Insights from GeophysicalData, Ophiolites and Oceanic Crust: New Insights from Field Studies and the Ocean
Drilling Program, 349 , 365–382, 2000.
Godfrey, N. J., and S. L. Klemperer, Ophiolitic Basement to a Forearc Basin and Implicationsfor Continental Growth: The Coast Range/Great Valley Ophiolite, California, Tectonics ,17 , 558–570, 1998.
Godfrey, N. J. and Beaudoin, B. C. and Klemperer, S. L. and Mendocino Working Group,Ophiolitic Basement to the Great Valley Forearc Basin, California, from Seismic and Grav-ity Data: Implications for Crustal Growth at the North American Continental Margin,Geological Society of America Bulletin, 109 , 1536–1562, 1997.
Herrmann, R. B., Computer Programs in Seismology: An Evolving Tool for Instruction andResearch, Seismological Research Letters , 84 , 1081–1088, 2013.
Jachens, R. C., R. W. Simpson, R. W. Graymer, C. M. Wentworth, and T. M. Brocher,Three-Dimensional Geologic Map of Northern California: A Foundation for EarthquakeSimulations and Other Predictive Modeling, Eos Trans. AGU , 87 , 2006, Fall Meet. Suppl.,Abstract S51B-1265.
Langenheim, V. E., R. W. Graymer, and R. C. Jachens, Geophysical Setting of the 2000 ML
5.2 Yountville, California, Earthquake: Implications for Seismic Hazard in Napa Valley,California, Bulletin of the Seismological Society of America, 96 , 1192–1198, 2006.
Lin, G., P. M. Shearer, E. Hauksson, and C. H. Thurber, A Three-dimensional CrustalSeismic Velocity Model for Southern California from a Composite Event Method, Journalof Geophysical Research: Solid Earth, 112 , B11,306, 2007, b11306.
Lin, G., C. H. Thurber, H. Zhang, E. Hauksson, P. M. Shearer, F. Waldhauser, T. M.Brocher, and J. Hardebeck, A California Statewide Three-Dimensional Seismic VelocityModel from both Absolute and Differential Times, Bulletin of the Seismological Society of
America, 100 , 225–240, 2010.
MacKevett, N. H., Structural Geology of the Sacramento Basin: Pacific Section, chap. TheKirby Hill Fault Zone, 61–78, Pacific Section of AAPG, 1992.
Moores, E. M., J. Wakabayashi, and J. R. Unruh, Crustal-Scale Cross-Section of the U.S.Cordillera, California and beyond, its Tectonic Significance, and Speculations on the An-dean Orogeny, International Geology Review , 44 , 479–500, 2002.
25
NCDEC, Northern California Earthquake Data Center, UC Berkeley Seismological Labora-
tory , Dataset , 2014.
Pepper, M. W., and D. S. Johnson, Structural geology of the Sacramento Basin: Pacific
Section, chap. The Midland Fault System, southern Sacramento basin, California, 27–40,Pacific Section of AAPG, 1992.
Rodgers, A., N. A. Petersson, S. Nilsson, B. Sjgreen, and K. McCandless, Broadband Wave-form Modeling of Moderate Earthquakes in the San Francisco Bay Area and PreliminaryAssessment of the USGS 3D Seismic Velocity Model, Bulletin of the Seismological Society
of America, 98 , 969–988, 2008.
Rojstaczer, S. A., R. E. Hamon, S. J. Deverel, and C. A. Massey, Evaluation of SelectedData to Assess the Causes of Subsidence in the Sacramento-San Joaquin Delta, California,Tech. rep., U.S. Geological Survey Open-File Report 91-193, 1991.
Sharma, P., C. E. Jones, J. Dudas, G. W. Bawden, and S. Deverel, Monitoring of Subsidencewith UAVSAR on Sherman Island in California’s Sacramento-San Joaquin Delta, Remote
Sensing of Environment , 181 , 218–236, 2016.
Thurber, C., and D. Eberhart-Phillips, Local Earthquake Tomography with Flexible Grid-ding, Computers & Geosciences , 25 , 809–818, 1999.
Thurber, C., H. Zhang, T. Brocher, and V. Langenheim, Regional Three-dimensional SeismicVelocity Model of the Crust and Uppermost Mantle of Northern California, Journal ofGeophysical Research: Solid Earth, 114 , B01,304, 2009.
Thurber, C. H., T. M. Brocher, H. Zhang, and V. E. Langenheim, Three-dimensional P-wave Velocity Model for the San Francisco Bay Region, California, Journal of Geophysical
Research: Solid Earth, 112 , B07,313, 2007.
Torres, R. A., N. A. Abrahamson, F. N. Brovold, L. F. Harder, N. D. Marachi, C. H. Neudeck,L. M. O’Leary, M. Ramsbotham, and R. B. Seed, Seismic Vulnerability of the Sacramento-San Joaquin Delta Levees, Tech. rep., Report of Levees and Channels Technical Team,Seismic Vulnerability Sub-team to CALFED Bay-Delta Program, 2000.
Wentworth, C. M., G. R. Fisher, P. Levine, and R. C. Jachens, The Surface of CrystallineBasement, Great Valley and Sierra Nevada, California: A Digital Map Database, Tech.rep., U.S. Geological Survey Open-File Report 95-96, 1995.
Wesling, J. R., and K. L. Hanson, Mapping of the West Napa Fault Zone for Input intothe Northern California Quaternary Fault Database, Tech. rep., US Geological SurveyExternal Award Number 05HQAG0002, 2008.
Wong, I. G., Structural Geology of the Sacramento Basin: Pacific Section, chap. EarthquakeActivity in the Sacramento Valley, California and its Implications to Active GeologicStructures and Contemporary Tectonic Stresses, 5–14, Pacific Section of AAPG, 1992.
26